Autonomous vehicles and mobility, in order to be reliable and efficient, rely on software architecture often composed of an initial component dealing with the perception of the road environment. This component must produce an accurate, up to date and reliable information on the positioning of the obstacles around the ego-vehicle. This work proposes an efficient way to address the first level with a two stages approach. The first stage is dedicated to the dynamic objects modeling, and the second one to the static scene modeling. This paper will focus on the second part by presenting a new concept of real time “continuous” obstacle occupancy map, based on the belief theory, that allow to manage road's static information involving large scale and non linear shapes (guardrail, safety wall, curved tunnel,…). This modeling is called Belief Plot Map (BPM). Some results with real data, different road configurations, in a global perception architecture (with the merge of the two levels: static and dynamic) will be presented in order to show the relevance and quantify the performance of this approach.
Belief Plot Map - An efficient way to model static objects and environments for Autonomous Driving Systems
2021-09-19
2104844 byte
Conference paper
Electronic Resource
English
European Patent Office | 2023
|METHOD FOR RECOGNIZING OBJECTS IN AUTONOMOUS DRIVING SYSTEMS
European Patent Office | 2024
|A Generalized Belief Space Approach for Autonomous Navigation in Unknown Environments
British Library Conference Proceedings | 2015
|SCENARIO IDENTIFICATION IN AUTONOMOUS DRIVING ENVIRONMENTS
European Patent Office | 2022
|